[lm-sensors] Possible support for the SMSC SCH5317? (found on HP dc7700 hardware)

[Jean Delvare]

Hi Juerg, hi Simon,

On Fri, 15 Jun 2007 09:08:59 -0700, Juerg Haefliger wrote:
> On 6/15/07, Simon Farnsworth <simon.farnsworth at onelan.co.uk> wrote:
> > Juerg Haefliger wrote:
> > > Simon,
> > >
> > > If you feel adventurous, you can add device ID 0x85 to the smsc47b397
> > > driver which also supports the sch5037 chip. You'd have to modify it
> > > as follows:
> > >
> > > static int __init smsc47b397_find(unsigned short *addr)
> > > {
> > >        u8 id, rev;
> > >
> > >        superio_enter();
> > >        id = superio_inb(SUPERIO_REG_DEVID);
> > >
> > > -        if ((id != 0x6f) && (id != 0x81)) {
> > > +        if ((id != 0x6f) && (id != 0x81) && (id != 0x85)) {
> > >                superio_exit();
> > >                return -ENODEV;
> > >        }
> > >
> > Did this, and the chip is found, but the numbers I get back look wrong:
> >
> > dmesg output from the module:
> > smsc47b397: found SMSC LPC47B397-NC (base address 0x0480, revision 2)
> >
> > sensors output:
> >
> > # sensors
> > smsc47b397-isa-0480
> > Adapter: ISA adapter
> > temp1:       +44°C
> > temp2:       +53°C
> > temp3:       +26°C
> > temp4:      -128°C
> > fan1:     2011 RPM
> > fan2:       82 RPM
> > fan3:     1262 RPM
> > fan4:     1095 RPM
> 
> Well I think that looks pretty good.

I agree it looks quite good. temp4 and fan2 are probably not connected,
thus the wrong values, but the 6 other values look just fine.

> > temp1 and temp2 both look likely for CPU, while temp3 might well be
> > system temperature. temp4 is completely out, and I'm not sure how I'd
> > relate the fan speeds to the fans in the system (I have a fan in the
> > front, a CPU fan, and a fan in the PSU.
> 
> Run some stress tests like 'stress' or 'cpuburn-in' to correlate temps
> to CPUs. Are the fans automatically adjusted? If yes, this should also
> tell you which one is the CPU fan, assuming it spins up under load.
> Alternatively, you can unplug one fan at a time to correlate the
> physical reality to the readings.

So it looks like we could easily add support for this chip. Juerg, can
you please take care of it? Pick list:

* patch for sensors-detect
* patch for the smsc47b397 driver
* update Devices in the wiki

Thanks,
-- 
Jean Delvare